Killer phones

Ten days ago Spain suffered its worst rail crash  since 1944, in Santiago de Compostella. 79 passengers died.

The crash took place on a new 90 km section of track between Santiago and Ourense. This is built for high-speed operation at 220 km/hr, but currently limited to 200 km/hr by signalling limitations. However, the accident was on a sharpish curve before the station, with a limit of only 80 km/hr. There is a single signal for deceleration.

Just before the accident, the driver took a mobile phone call from the ticket collector, requesting an additional stop before Corunna (and well beyond Santiago) to let off a family of passengers. There was no urgency for the call. [Update from comments:> the handset may well have been a rail=specific 2-way radio, not an ordinary mobile, but it makes no difference to my argument.]

The magistrate investigating legal liability for the accident – there is of course another technical investigation under way – has blamed the driver for going too fast against signals and the route plan. He has not given weight to the phone call; it’s unnecessary to fix the driver’s responsibility, and perhaps he wanted to avoid the impression of blaming the unfortunate but surely innocent ticket collector and the family he was trying to help.

However, common sense suggests the phone call was significant. The driver was experienced and not fatigued – he only took over at Ourense, and knew the track well. The weather was clear. There does not seem to have been any equipment failure (as opposed to design flaws). The phone call is the standout differentiating factor. There’s a mass of evidence that phone calls are distracting to car and truck drivers and responsible for a good many vehicle accidents and deaths.

The USA lags well behind international practice in legislating against the use of mobile phones at the wheel. 66 countries, covering the great majority of the world’s population, have bans. Only seven US states are reported by Wikipedia as having general bans (California, Connecticut, Delaware, DC, Hawaii, Illinois, New Jersey). Several others have footling partial bans on learner drivers, at night, near schools and so on.

Contrary to intuition and folk wisdom, hands-free phones make little difference. It’s the speech that’s distracting, not the use of one hand.

Just turn it off.

***************************
PS on signalling
For high-speed rail buffs only. The track section is designed and equipped for the latest ERMTS in-cab signalling which among many other goodies brakes the train automatically if speed limits are exceeded. That’s the system in use on the true high-speed AVE trainsets on the lines going out from Madrid to Barcelona and Seville. But the line doesn’t run all the way to Galicia, and the section to Ourense is still under construction. The short high-speed track to Ourense is a vanity regionally-driven anticipation, using Iberian gauge. The ALVIA trainsets like the one that crashed are expensive electric/diesel hybrids with dual-gauge bogies (not SFIK implicated). The manufacturers of the trainsets and the signalling between them have not been able to get the complex ERMTS software to work properly. (Source: El Pais.) For now the ALVIAs have to use the older and less capable ASFA system – which does not include automatic braking. It’s rated up to 200 km/hr, but in hindsight this looks a mistake, along with the single trackside signal to announce the drastic speed reduction.

Comments

  1. Warren Terra says

    (Edit)

    Your postscript seems to indicate that there’s supposed to be an automated imposition of the speed limit, but they haven’t gotten it to work. It’s good to hear it’s supposed to be there, but this really does seem rather inexcusable; I realize this can be a complicated thing to implement if it’s tying into some whole set of route and train management software, but dumb, simple fixes to just enforce or just loudly warn of speed limits could pretty easily be made, and made foolproof, at minimal cost compared to the track construction.

    • James Wimberley says

      (Edit)

      “Dumb, simple fixes” to hugely complicated software suites? Hum.

      I agree that the failure to fix, after a long delay, is inexcusable. Spain is a crony capitalist government; you don’t see fines for delays in building motorways, or prison for corrupt bankers, either.

      • Warren Terra says

        (Edit)

        No; dumb, simple fixes to train speed, in parallel to unworkable hugely complicated software suites. You could work something up using technology most of a century old that would let off a loud alarm when the train is much too fast for the section of track; slightly smarter options are available.

      • Mike says

        (Edit)

        James wrote:
        “Spain is a crony capitalist government; you don’t see fines for delays in building motorways, or prison for corrupt bankers, either.”

        Sounds a lot like the good ol’ USA to me, although the occasional contractual monetary “penalty” for failure to complete a road project in timely fashion does creep in from time to time…

    • Mike says

      (Edit)

      Warren,
      Yes, a bit confused description…basically the crash occurred near where the transition from where the full signal system with auto-stop features is in operation and the older, non-auto-stop was still in place.

      BTW, having a single signal to protect/restrict track for speed or other reason is not unusual. In retrospect, it may prove that was inadequate if it turns out the signal was somehow not obvious enough for crews to notice.

      There’s another factor I don’t think has been mentioned. The engineer was only recently qualified to work on this section of track. That means he’s been trained well enough to operate on his own in that territory, but his lack of experience and unfamiliarity with that transition between signals systems and the speed restriction may nonetheless have contributed to the accident.

      • James Wimberley says

        (Edit)

        “..the crash occurred near where the transition from where the full signal system with auto-stop features is in operation and the older, non-auto-stop was still in place.”
        I trust El Pais and Spanish Wikipedia here. The design is for ERMTS (with automatic braking) to operate to within a few km of the station, then hand over to AFSA for the low-speed arrival. But since ERMTS doesn’t work, they were using ASFA all the way. ERMTS would have started braking and given warnings of the handover.

  2. prognostication says

    (Edit)

    Pedantic note: There are general bans on driver cell phone use in more states than those listed, but it’s a secondary offense in many of them, meaning the police have to stop a driver for something else to ticket them for cell phone use. That is the law in my current state of residence, as I understand it, and several others in this region.

    • says

      (Edit)

      Wikipedia puts only Maryland in that box, though I’m sure you are right and they wrong about others. But secondary bans fall into my “footling” category, the pretence of action instead of the real thing.

      • Henry says

        (Edit)

        And Maryland limited the ban, such as it is, to hand-held phones, even though it was quite clear that the use of non-hand-held phones was equally dangerous. Maybe the companies that make non-hand-held phones were more generous in their campaign contributions; I can’t think of any other reason why legislators would want to allow more people to die in accidents.

  3. Herschel says

    (Edit)

    I entirely agree that the problem with phone calls while driving is the conversation, not the use of a hand. One of the very worst driving errors I’ve ever made, which happily had no adverse consquences, occurred because I was engaged in conversation with a passenger sitting next to me. (For those who know Washington, I was turning right on to Rock Creek Parkway from Virginia Avenue. This was part of my daily homeward commute at the time, and during evening rush hour you don’t have to stop before making the turn. I had stayed late at my office, however, and it was no longer rush hour, but being engaged in conversation, I was driving, as it were, on autopilot. I breezed through the red light. Luckily, nobody hit me. It might very well have been fatal to me if they had.)

    • Henry says

      (Edit)

      I’ve read that conversations with passengers are much less dangerous than phone conversations, even with non-hand-held phones, because a passenger can see when the driver needs to concentrate on his driving and the passenger will stop talking during that time.

      • Herschel says

        (Edit)

        I can see how that might well be true. In my little story, however, I was doing all the talking at the crucial moment.

    • Rud Merriam says

      (Edit)

      More directly apropo to your comment is the use of ham radios in vehicles. These usually involve the use of a hand held microphone – or even a telegraph key strapped to the leg. But studies indicate (sorry, no citation available) there is NO increase in driving problems for amateur radio operators.

      One possible reason is that people are trained to consider the telephone conversation as important. Thus, people focus on the conversation. Ham operators, in opposition, know their conversation is casual so do not focus as much.

      There may also be an effect due to the implied expectation of the other party in the conversation. Hams know and respect that a ham in a vehicle may be interrupted by the requirements of driving. There is no expectation of undivided attention. Here again the training for phone conversations is one of undivided attention.

      Finally, ham radios are semi-hands free since the speaker is on the radio, not a handset. When I am listening to the other party or require both hands on the wheel I simply put the microphone in my lap.

  4. NCG says

    (Edit)

    Maybe we should go ahead and actually just confiscate the phones beforehand. I’d rather do that than crucify people afterwards. You could still have an intercom for *emergencies*. And I think I still prefer having a human in charge than a machine, though if the machine just prompted the person, or reminded them, to slow down, that would be okay.

    • James Wimberley says

      (Edit)

      Several cities have the driverless French-designed VAL light metros, which work fine and have SFIK a very good accident record. There will soon come a point when the human controller of any vehicle is seen as on balance a safety liability.

  5. H says

    (Edit)

    Yep, it is true that talking on your cell phone while driving is illegal in California.

    But, no one seems to care. Certainly, not the police.

    I see dozens doing it every day.

    Every time I call AT&T customer service I do get a hectoring message about not texting and driving. Kinda weird when I am always calling about a landline or ‘net service, not cell.

  6. says

    (Edit)

    I find it interesting that you seem to combine two very separate thoughts into one in this post. The first thought is that cell phone use is a dangerous distraction, don’t do it (just turn it off). The second thought is that legislation is needed. Yet no attempt is made to connect the two.

    This is just an observation, not necessarily a criticism — after all, you may have independently considered the evidence and concluded that legislation is the proper course and just felt that this wasn’t the post to share that information. It seems to me that a proper intellectual analysis requires that we establish 1. that legislation would solve or significantly reduce the problem, and 2. that legislation is the only, or at least the best, solution to do so. Who knows? Maybe education or peer pressure would be more effective. Or maybe legislation is actually the best approach.

    This is such a common thing we do. We have a strong tendency to operate on the horribly flawed “this is bad: therefore, legislation” syllogism. It’s given us decades of drug war, endless fights over abortion and all sorts of other societal problems. And legislation we do pass oftentimes ends up failing to fix, or even exacerbates, the problem.

    It would be nice if we took more time to first ask the question, “Is this problem best served by legislation?”

      • says

        (Edit)

        Wow.

        We really have sunk low, if the actual question can’t even be discerned. Thanks for that perfect example.

        In case it isn’t clear, the actual question is: “How do you propose to have people stop using cell phones while driving?” And no, that is not the same question.

        • James Wimberley says

          (Edit)

          The costs of a ban are very low, and the conduct stigmatized is clearly dangerous to third parties. A low success rate would still meet a cost-benefit test. I think the onus of proof is on the opponents of legal bans.

          • says

            (Edit)

            If the low success rate were lower than following some other course of action (not saying there is one, just pointing out the analysis needed), then the bad could be higher in cost.

            Again, you may be right, but I find the notion of the “onus of proof on the opponent of bans” to be a somewhat bizarre approach (and dangerous, given legislators’ desire to get their names on bills).

            One of the dangers of putting faith in bans is the sense that once you’ve banned something, you’ve solved the problem. As we know, that is never true. In places where talking on cell phone is banned, for example, it’s very hard to enforce, whereas campaigns that get passengers to refuse to allow the driver to talk on the phone may be more effective. I’m just saying, “let’s have the discussion before we ban.”

          • A Critic says

            (Edit)

            “The costs of a ban are very low, and the conduct stigmatized is clearly dangerous to third parties.”

            The costs of a ban include liberty, the most precious thing in the known universe, so the costs are very high.

            If the conduct was really dangerous then there would be hundreds of millions of cell phone caused accidents every week.

        • J. Michael Neal says

          (Edit)

          In case it isn’t clear, the actual question is: “How do you propose to have people stop using cell phones while driving?” And no, that is not the same question.

          We disagree on your last sentence. Absent a ban you might reduce people using cell phones while driving but you will not stop them. If that’s what you want to do, a ban is your only option.

          • says

            (Edit)

            And you are assuming that with a ban you will stop them. Is there any evidence that a ban actually stops the activity? Reduces it maybe, but I’ve seen no evidence that a ban will stop people using cell phones. So now we’re back to analyzing what option works best to maximize benefit, and it isn’t necessarily automatically a ban.

  7. Curmudgeon says

    (Edit)

    With regard to the train crash — I think you’re too caught up on reports that the driver was using a cell phone.

    Most trains have radios. This train used a cell-phone radio, most likely based on GSM-R. If the train didn’t have a GSM-R radio, it would have had a push-to-talk radio. Either way, the driver would have been told to make the station stop verbally. I don’t see a GSM-R radio as inherently more distracting than a two-way radio.

    For decades, train operations have relied on verbal orders given via radio. Being able to take radio calls–be they GSM-R or two-way–while operating a train is an essential part of the driver/locomotive engineer’s job, in just the same way that being able to take radio calls while flying is an essential part of an aircraft pilot’s job.

    The cell phone isn’t the problem here.

    • Mike says

      (Edit)

      Curmedgeon is correct. The phone was a work-related call on a work phone, so I presume needed to be answered by the engineer. It’s like any other distraction in the cab, if has to be managed by the engineer…

      That said, the engineer is ultimately responsible for the safe operation of the train and that definitely includes observing signals or other restrictions. That was the failure here, not in answering or talking on the phone.

      The accident occurred near where the non-operational auto stop feature would have otherwise come into play. This issue and the phone means it will be a very difficult situation to unravel in terms of what could have been done to prevent it…but one thing is certain, the engineer failed in his duty to properly prioritize with safety being the first priority in railroading.

      • MobiusKlein says

        (Edit)

        It’s a system with a single point of failure, and that point of failure was a human.

        Should there be two engineers there, or otherwise empowered rail-dudes?

        • Mike says

          (Edit)

          Mobius,
          Maybe, maybe not. A second human was standard in loco cabs until steam passed mostly into history, because of the need for a fireman. It was standard practice in many cases for one crew member to call the signals and the other to verbally confirm the sighting, i.e. Eng: “Green.”
          Fireman: “Green!”

          In modern times, the fireman is gone in most North American RR practice. Most crews are down to two people, an engineer and a conductor, who rides up front now that cabooses are gone, too. So there is still a human backup in many cases on freight lines. Passenger lines often have only one crewmember up front, because the conductor is back in the train working the passenger cars. It’s also the case that two people can end up with one distracting the other. And even with two people…

          A friend used to be in crew service with one of the major lines as a conductor. Early one morning, they both were apparently aslep when they ran a red signal and T-boned another RR’s train where it crossed at grade on a diamond. Fortunately, all escaped without serious injury, but there was a fire and the usual damage in the millions of dollars.

          Yes, humans err, but there are continual efforts to try to provide improvements. Not all are supported by the RRs, in fact, many of the most important ones were forced on them by government regulation after the RRs failed to act on them for reasons of cost, etc. The latest in the US is what’s called positive train control (PTC) where each unit continually broadcasts its location and a central computer ensures anticipated movements will not come into conflict. Sort of see-and-avoid on cell phones, but that’s a vast oversimplification so don’t take the analogy too far. FRA wants it sooner, the RRs are resisting because of cost, supposedly lack of technical maturity, etc. All I can say is if it works for collision avoidance for planes — a very similar app in use for years now — it’s probably OK for much slower moving trains.

          • MobiusKlein says

            (Edit)

            I was specifically thinking of the Spanish crash case, where there were no automatic systems to control the train. In that specific case, I certainly would expect some other factor to be present. And until we can get those automatic controls, a second human is the best idea I have.

            I’m pretty sure the Caltrain system in the SF Bay Area has more than one controller on trains, in addition to conductors. Might be because they run the trains in reverse going from San Jose to SF and need a spotter in the ‘back’.

    • James Wimberley says

      (Edit)

      You are agreeing with the investigating magistrate, and in terms of professional responsibility, that’s correct.

      However there are several features of the situation that imposed an extra cognitive load on the driver.
      (1) As I read the reports, the call was not an instruction (as it would be from a control centre) but a request. The driver therefore had to make a decision, which is harder than following an order.
      (2) The deceleration was extreme – 200 to 80 km/hr – and notified by a single trackside signal. Miss that, and you are in big trouble.
      (3) The section has a large number of tunnels, all cognitively the same. I read some suggestion that the driver did not know where the train was (which ERMTS would have told him).

      The form of the device does not matter, only the distraction.

  8. ferd says

    (Edit)

    I’d want to know who, exactly, made the genius decision to set up JUST the “single trackside signal to announce the drastic speed reduction.” That person can’t be allowed to make any more big decisions.

    I’d want to know if ANY of the train drivers who drove past this single trackside genius-signal ever thought, “Man, just this one stupid little signal to decelerate! This is gonna cause an accident!!” I’d like to know which, if ANY, of the drivers ever passed along a suggestion to somebody up the chain, that the signal was an accident waiting, and something needed to be done in the interval before the fancy automated tech took hold.

    I’d want to ask a junior high school iPhone app. developer how long it would take her to gin up some code for an iPhone so the iPhone, through GPS, knew where the train was, how fast it was moving, and whether the iPhone should yell out loud to slow down the dang train.

    This is in the class of Palm-To-Face incredible as the stories of Iraqi families shot and killed at military checkpoints because they didn’t stop their cars at checkpoint signs that said “STOP!” . . . in english. As in, did any of this junk REALLY happen? Were the signs REALLY just written in english? VERY hard to believe the Iraq story, and this train genius-single-signal story.

    • James Wimberley says

      (Edit)

      The trackside signals are a stopgap and fallback. ERMTS doesn’t need them. I should have mentioned that ERMTS works fine on a good number of other high-speed lines, such as the French LGV-Est to Metz (320 km/hr) and the Dutch-Belgian HSL-Zuid.

  9. ferd says

    (Edit)

    I’d want to find out if any of the train drivers wanted to alert higher ups to the single-signal problem, but were afraid to do so. And if they were afraid, I’d want to find out why, and fix that if possible. Is there somebody in power over there who made drivers afraid to raise a red flag?

  10. Matt says

    (Edit)

    If, as you say “speech is the problem”, why is talking to a friend in the car not as dangerous as talking on a cell phone? And, if it is, why would we not ban passengers in vehicles (or at least talking to passengers)?

    I’m not for mixing cell phones and cars, but this seems to be the obvious conclusion to your argument.

    • NY-Paul says

      (Edit)

      “…..why is talking to a friend in the car not as dangerous as talking on a cell phone? “
      ——————————————————————

      It’s a question of degree. Talking to a friend in the car CAN BE as dangerous as talking on a cell phone if both parties are irresponsible jerks and permit the conversation to dangerously distract the driver from his/her primary responsibility. However, I would suggest that a passenger engaged in a conversation with the driver is, first of all, physically right next to the driver and can assess how attentive the driver is to his/her driving. He/she can take corrective measures should the driver wander dangerously away from what should be his/her primary focus. As the term, “back seat driver” indicates, normally, when you have a driver and a passenger, both parties are, to one degree or another, involved in driving the car.

      Also, to suggest (not you, by the way) that a pilot, and/or engineer talking to a second party via a phone-like device is somehow analogous to a civilian talking on a cell phone while driving a car is a mistaken correlation. The question here is, what is the focus of the conversation? We can assume that with a pilot/engineer the conversation is with another professional, air traffic controller for instance, and the focus of the conversation is in cooperating, and sharing information for the single purpose of operating the equipment involved. A driver talking on a cell phone, however, is operating at cross purposes with the other party. The driver should be concentrating on driving, but the other party is mentally engaging his/her mind with issues that have nothing to do with driving.

      • James Wimberley says

        (Edit)

        On your last paragraph: the air traffic controllers and pilots are generally talking in a stylised idiolect. The conversation should not generally include surprises. Sometimes it does; say a passenger has had a heart attack. The train situation is similar to cabin crew telling the pilot about a heart attack while she is on approach. They wouldn’t do it; the cabin crewperson, like the car passenger, has information about the context that inhibits the distraction. Ticket collectors on trains have only a general sense where the train is, and no knowledge of signals.

Trackbacks